Background: A protocol for the contouring of target volumes in lung cancer was implemented. Subsequently, a study was performed in order to determine the intra and inter-clinician variations in contoured volumes. Materials and methods: Six radiation oncologists (RO) contoured the gross tumour volume (GTV) and/or clinical target volume (CTV), and planning target volume (PTV) for three patients with non-small cell lung cancer (NSCLC), on two separate occasions. These were, respectively, a well-circumscribed T1N0M0 lesion, an irregularly shaped T2N0M0 lesion, and a T2N2M0 tumour. Detailed diagnostic radiology reports were provided and contours were entered into a 3D planning system. The target volumes were calculated and beams-eye view (BEV) plots were generated to visualise differences in contouring. A software tool was used to expand the GTV and CTV in three dimensions for an automatically derived PTV. Results: Significant inter-RO variations in contoured target volumes were observed for all patients, and these were greater than intra-RO differences. The ratio of the largest to smallest contoured volume ranged from 1.6 for the GTV in the T1N0 lesion, to 2.0 for the PTV in the T2N2 lesion. The BEV plots revealed significant inter-RO variations in contouring the mediastinal CTV. The PTV's derived using a 3D margin programme were larger than manually contoured PTV's. These variations did not correlate with the experience of ROs. Conclusions: Despite the use of an institutional contouring protocol, significant interclinician variations persist in contouring target volumes in NSCLC. Additional measures to decrease such variations should be incorporated into clinical trials.